[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <200612101109.34149.stefan@loplof.de>
Date: Sun, 10 Dec 2006 11:11:32 +0100 (MET)
From: Stefan Rompf <stefan@...lof.de>
To: Thomas Graf <tgraf@...g.ch>
Cc: David Miller <davem@...emloft.net>, drow@...se.org,
dwmw2@...radead.org, joseph@...esourcery.com,
netdev@...r.kernel.org, libc-alpha@...rceware.org, akpm@...l.org
Subject: Re: [NETLINK]: Schedule removal of old macros exported to userspace
Hi,
Am Sonntag, 10. Dezember 2006 00:28 schrieb Thomas Graf:
> I had a short glance at dhcpclient's netlink code and found half
> a dozen bugs within five minutes all caused by cut and pasting
> sniplets from other netlink code.
Please send me the list of bugs you've spotted. Of course I want to fix them.
> It is my opinion that people
> will continue to drag along bugs this way unless the interface is
> removed. I never believed in providing backward compatibility to
> broken code.
My main problem with netlink was missing or wrong documentation (and yes, I
know it's much easier writing code than writing docs, especially if OSS
development is just a hobby). F.e. until 2.6.19 it has not been possible to
query one ifi_index with RTM_GETLINK even though rfc3549 specified this
operation. Instead, the application had to dump all interfaces and filter the
record it is interested in. Stuff like this creates much more bugs than a
non-typesafe macro because it makes you think "Damn it's working now, never
touch that code again".
Don't take this as an offense, I do value your work and will use libnl for
future projects, but I definitely do not share your opinion on how to deal
with ugly/obsolete userspace interfaces (even though this one is not
obsolete, because it is the only low level interface that exists at all).
Stefan
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists